Biological Analogues. Synthetic Iron(II1) - American Chemical Society

Swaddle, T. W.; King, E. L. Inorg. Chem. 1964, 3, 234. Espenson, J. H. Chemical Kinetics and Reaction Mechanisms;. McGraw-Hill: New York, 1981; pp 69-...
0 downloads 0 Views 637KB Size
Inorg. Chem. 1987, 26, 396-400

396 the sequence (with charges omitted)

-

2Cr(OH2)4(nic)2

Cr(OH2)3(nic)3

k21

Cr(OH2)5nic

um(II1) ion is approximately 20-fold greater than that for aquation of Cr,(nic)OH4+.) Thus, the spectral changes occurring during the observation period are due to the reactions

-

kl with k3, = 1.44 X lo4 s-I and kzl = 3.45 X 10" s-l. (An alternate Cr2(OH,),gicS+ f H 2 0 Cr(OH2)63+ Cr(OH2)5nic2+ i n t e r p r e t a t i ~ nin~ which ~ k z l > k 3 , does not seem reasonable.) With these values for the two rate constants, the maximum relative k, Cr(OH,)?+ Hnic Cr(OH2)5nic2+ H 3 0 + concentration of bis species, which occurs at -440 min, is -0.91.30 The rate of aquation of the dimer Cr2(nic)OH4+in perchloric The independently determined values of k2 for this temperature acid was determined at 60.0 "C in an experiment with [Cr(III)] can be extrapolated at this acidity; k2([H'] = 0.050 mol L-], 60 = 1.54 X mol L-' and [H'] = 0.05 mol L-l. Observations "C) = 1.3 X s-]. If this value is used, the value of k l can were made at t = 0 and from t values from 23 h to 11 days; during be derived from the spectral data. In this data treatment, the value this time, the species is aquated completely to hexaaquachroof the absorbance due to Cr(OH2)snic2+,the reactant in this k2 mium(II1) ion with a decrease in light absorption at 570 nm of step, must be assumed. Rather than use the value determined 18%. For comparison a solution of hydrolytic dimer with charge under different conditions, we assumed various values between 4+, described by Thompson et al.,19 was treated similarly. Plots the initial value and that due to hexaaquachromium(II1). It was of log ( A , - A,) vs. time for both species were approximately linear found that the value of k2 was not particularly sensitive to the over the course of the study (87% completion), suggesting that assumed value for this parameter. The derived value of k2 changes one of the steps in the aquation mechanism is slower than the by -6% if the assumed value for the light absorption due to others. From these data the empirical rate constants for aquation Cr(OH2),nic2+is changed by -5%. (As already mentioned, the of the nicotinato dimer and hydrolytic dimer are 2.7 X 10" and total change of light absorption was only 18%.) For the as6X s-*, respectively. sumption that light absorption due to equal amounts of CrThus, the empirical rate constant for aquation of Cr2(nic)OH4+ (OH2)5nic2+and Cr(OH2)63+was midway between that for no is appreciably smaller than that for the production of C T ( O H ~ ) ~ ~ +aquation (Cr2(OH2)lonic5+) and 100% aquation (Cr(OH,)63+), either from Cr(nic)2+ or from the hydrolytic dimer. For Cr2s-l. (With k , = 2.8 X the derived value of kl is 2.8 X (nic)OH4+ with the structure already proposed, it is reasonable s-l and k2 = 1.3 X s-l, one can calculate that during the to assume that breaking of the hydroxo bridge occurs rapidly on aquation the maximum relative concentration of Cr(OH2),nic2+ the time scale for breaking the nicotinate bridge. (The rate is only 7%. Thus, it is reasonable that the value of k , derived by constant for the conversion of hydrolytic dimer to aquachromiconsidering the sequence of two reactions is very close to that obtained simply as k = -A[ln ( A - Am)]/At.)

+

-

+

+

-

-

Yatsimirskii, K. B.; Federova, T. 1. Zh. Neorg. Khim. 1956. I , 2310. Swaddle, T. W.; King, E. L. Inorg. Chem. 1965, 4 , 532. Swaddle, T. W.; King, E. L. Inorg. Chem. 1964, 3, 234. Espenson, J. H. Chemical Kinetics and Reaction Mechanisms; McGraw-Hill: New York, 1981; pp 69-71. Although Cooper et a1.I" state that in 2 M HNOl the %sible spectrum of bis(nicotinato)chromium(III) showed no deterioration with time", no details regarding the duration of the observations are given.

Conclusion It is clear from this work that there are several inert oxygenbonded (nicotinato)chromium(III) species. The ion-exchange procedures employed in the present study provide the basis for the preparation of solutions of relatively pure species. The role of each of these in glucose tolerance can be tested.

Contribution from the Lash Miller Chemical Laboratories, University of Toronto, Toronto, Ontario, Canada M5S 1Al

Biological Analogues. Synthetic Iron(II1)-Specific Chelators Based on the Natural Siderophores Mark S. Mitchell, Debbie-Lee Walker, John Whelan, and B. Bosnich* Received July 11, 1986 Two synthetic analogues of the the naturally occurring iron(II1)-specific siderophores, one with catechol groups and the other with hydroxafic acid groups,have been prepared. Both are sexidentate chelates consisting of three catechol and three hydroxamic acid groups attached to a triamine "platform" via amide linkages. Both ligands are specific for iron(II1) and form very stable iron(II1) complexes; the stability constant for the iron(II1) tris(hydr0xamate) complex was determined to be about 28.

HemochromatosisiS2 refers to a group of disorders that result from a progressive increase in deposition of iron in the parenchymal and reticuloendothelial cells of the liver, pancreas, heart, and pituitary and that lead, successively, to tissue damage, functional insufficiency of the organs, and eventually death. Idiopathic hemochromatosis is an inherited autosomal recessive trait that manifests itself after the breakdown of the regulatory mechanism for iron absorption. The gene frequency for certain ethnic groups can be high, possibly of the order of 1 in 40 of the population, (1) Bothwell, T. H.; Charlton, R. W.; Cook, J. D.; Finch, C. A. Iron Metabolism in Man; Blackwell: London, 1979; p 105. (2) Fairbanks, V. F ; Fahey, J. L.; Beutler, E. Clinical Disorders of Iron Metabolism; Grune and Stratton: New York, 1971; p 399.

0020-1669/87/1326-0396$01.50/0

but the inherent propensity for the disorder appears to be triggered by environmental factor^.^ There are epidemiological associations with alcoholism, and excessive iron intake over a long time, particularly in combination with alcohol (viz. Bantu siderosis) .4 The life expectancy of untreated patients is about 4 years after the disease has become clinically manifest. Idiopathic hemochromatosis is now treated by phlebotomy in combination with chelation therapy. (3) Isselbacher, K. J., Adams, R. D., Braunwald, E., Petersdorf, R. G., Wilson, J. D., Eds.; Harrison's Principles of Internal Medicine, 9th ed.; McGraw Hill: Toronto, 1980; pp 488, 1552. (4) Bothwell, T. H.; Seftel, H.; Jacobs, P.; Torrance, J. D.; Baumslag, N . Am. J . Clin. Nutr. 1964, 14, 41.

0 1987 American Chemical Society

Synthetic Iron(II1)-Specific Chelators

Inorganic Chemistry, Vol. 26, No. 3, 1987 397

HN'

"H'

2

N

Figure 2. Structural formulas of the two siderophore analogues, 1 and 2.

Figure Structural formulas of three siderophores: ( , enterobactin (enterochelin); (B) ferrichrome family; (C) desferrioxamine B.

a%

Not surprisingly, particular attention has focused on the strong iron(II1)-specific chelators produced by microorganisms, the siderophores,5 to mobilize the tissue-deposited iron. The siderophores, in the main, employ catecholate and hydroxamate ligands for chelation to iron(II1). Three examples are shown in Figure 1, including desferrioxamine B, one of the chelators currently used in therapy. This chelator is administered subcutaneously and causes acute discomfort. Accordingly, various attempts have been made at developing more effective siderophore analogue^.^-^^^^.^^ This paper describes the synthesis of two sexidentate siderophore analogues, one a catechol and the other a hydroxamic acid ligand. The two target molecules are shown in Figure 2. Although the

-1

4

i WBULI/TME~

2'co2

1

p

HO ' 0

"!

soc12 J

( 5 ) Keller-Schierlein, W.; Prelog, P.; Zahner, H. Fortschr. Chem. Org. Naturst. 1964, 22, 279. Neilands, J. B. Struct. Bonding (Berlin) 1972, 11, 145. (6) Raymond, K. N.; Carrano, C. J. Acc. Chem. Res. 1979, Z2, 183 and

references therein. (7) Sun, Y.; Martell, A. E.;Motekaitis, R. J. Inorg. Chem. 1985,24, 4343. ( 8 ) Lee, B. H.; Miller, M. J.; Prody, C. A.; deilands, J. B. J. Med. Chem. 1985, 28, 317, 323. (9) Weitl, F. L.; Raymond, K. N. J. A m . Chem. SOC.1979, 101, 2728. Carrano, C. J.; Raymond, K. N. J. A m . Chem. SOC.1979,101, 5401. (IO) Venuti, M. C.; Rasstetter, W. H.; Neilands, J. B. J . Med. Chem. 1979, 22, 123. ( 1 1) Pecoraro, V. L.; Weitl, F. I.; Raymond, K. N. J . Am. Chem. SOC.1981, 103, 5133. (12) Abu-Dari, K.; Raymond, K. N. J . Am. Chem. SOC.1977, 99, 2003. (13) Harris, W. R.: Raymond, K. N . J . A m . Chem. SOC.1979, 101,6534. (14) Anderegg, C.; I'Eplattenier, F.: Schwarzenbach, G. Helu. Chim. Acta 1963, 46, 1409. (15) Perrin, D. D.; Dempsey, B. Buffers f o r p H and Metal Ion Control; Chapman and Hall: New York, 1983. (16) Schwarzenbach, G. Complexomerric Titrations; Interscience: New York, 1957. (17) Schwarzenbach, G.; Schwarzenbach, K. Helu. Chim. Acta 1963, 46, 1390. (18) Cooper, S. R.;McArdie, J. V.; Raymond, K. N. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 3551.

TWW 1

N

I NY/CH~OH

.1

2

N

Figure 3. Outline of the synthesis of the ligands 1 and 2

catechol analogue has been synthesized before:,'0 we describe here a different method that, arguably, has some advantages over the previous strategies. Synthesis The synthesis of the siderophore analogues is outlined in Figure 3. One of the obvious, direct methods of synthesizing the triamine

Mitchell et al.

398 Inorganic Chemistry, Vol. 26, No. 3, 1987 platform 11, by (BH,) reduction of the corresponding triamide (which is readily obtained from the commercially available tris(acid chloride)), proved unsatisfactory. The triamide is an extremely insoluble solid that is difficult to manipulate and that is very slowly reduced by BH3 in T H F to give, after acidification, the trihydrochloride salt of 11as a sticky solid that proved difficult to purify. This method was employed in the previous s y n t h e s i ~ . ~ By contrast, the N-methyl triamide is easy to handle, and we found that it is readily reduced with BH3.S(CH3)2in T H F to give a good yield of the trihydrochloride salt as a stable crystalline solid.” This tri-N-methyl analogue of 11, however, proved unsatisfactory as a platform because of the decreased rate of amide bond formation in certain cases over that observed for the primary triamine platform. We therefore chose the classical conventional route to 11 (Figure 3 ) , which, without difficulty, gave the desired triamine platform in satisfactory yield. The activated and protected catecholate arm 6 was obtained from catechol by the conventional means outlined in Figure 3. Reaction 4 5 proved troublesome, giving erratic and generally low yields. Coupling of 6 with platform 11 was most expeditiously accomplished under Schotten-Baumann conditions using a water-methylene chloride dispersion. The product, 17, is a readily crystallizable solid. Deprotection of the catecholate acetonide was accomplished at 25 “ C in trifluoroacetic acid-water solution; the ratio of the solvents was adjusted so that the enterobactin analogue 1 precipitated from solution as the reaction progressed. The synthesis of the hydroxamic acid arm began with the (Michael) addition of (E)-benzaldoxime (12) to methyl acrylate to give the nitrone 13. Base hydrolysis to remove the ester group, followed by the hydrazine induced removal of the benzylimine group, gave 3-(hydroxyamino)propionic acid (14). Reaction of 14 with benzoyl chloride under Schotten-Baumann conditions gave the protected hydroxamate 15, the carboxylate group of which was activated with DCC to give the p-nitrophenyl ester 16 as a highly crystalline solid. Platform 11 and hydroxamate arm 16 were coupled .in THF-H20 in the presence of triethylamine to give 18. Deprotection of 18 with methanolic ammonia at 25 “ C gave the desired product 2 as a crystalline solid. The synthesis of the hydroxamate arm does not appear to be a general strategy; it seems to be specific for benzoyl or other aromatic acyl derivatives. Thus, for example, attempts at preparing 15 with acetyl rather than benzoyl groups failed by the procedures given here. We had attempted to use the nitrone, 13, as a protecting group, but the corresponding acid of 13 could not be activated for amide formation without causing Beckmann rearrangement (to the benzamide) at the nitrone functionality. Similarly, we were unable to prepare the general system 19, an active ester that could potentially provide a direct, simple, general method of coupling of the arms to platform 11 to give a variety of derivatives.

I I’

R 19

Iron(II1) Complexes When the tris(catecho1) ligand 1 is allowed to react with iron(II1) in solutions maintained at pH values greater than 8.5, the violet tris(catecho1ate) complex is fully formed. Similarly, the tris(hydroxamic acid) ligand 2 gives the fully formed teacolored iron(II1) tris(hydroxamate) complex under the same conditions. The latter complex, however, is insoluble in all common solvents except for D M F and Me2S0, and hence we resorted to the use of solvent mixtures for subsequent work on this complex. We chose 3:l DMF-water mixtures buffered with borax at pH 8.5. For this mixed-solvent buffer, the solution electronic absorption spectra of the complex [Fe(2’)I0 and of [Fe(l’)l3- in 98% water-methanol at pH 8 . 5 , where 2’ and 1’ refer to the fully



! i

I

1

-

0

I

i

1

0

1

1 300

1

500

400

7 0

600

nm-+

Figure 4. Visible absorption spectrum of [Fe(2’)] in 3:l DMF-water solution at pH 8 . 5 .

I

01

i 303

I

400

1

500

I

600

700

1

803

nm+

Figure 5. Visible absorption spectrum of [Fe(l’)13-in 98% watermethanol solution at pH 8.5.

deprotonated ligands 2 and 1, respectively, are shown in Figures 4 and5. The spectrum of [Fe(1’)I3- is identical with the one described previously: and that of [Fe(2’)J0is very similar to those observed for the naturally occurring iron(II1) tris(hydr0xamate’) comp1exes.l2 The (Ph,As),[Fe(l’)] salt and the neutral [Fe(2’)] compound could be isolated as deep violet and brick red powders, respectively. In neither case could we obtain crystals suitable for X-ray analysis. The stability constant of [Fe(1’)I3- has been determined in water solution; a value of about 46 was reported.13 As expected, the catechol ligand 1 forms much stronger complexes with iron(1II) than does the hydroxamic acid ligand 2. Thus we find that, in our 3:1 DMF-water pH 8.5 buffer, ligand 1 fully displaces the hydroxamate ligand from [Fe(2’)] after 10 h at 25 OC when M equimolar amounts of 1 and [Fe(2’)] are used at 2.5 X concentrations. In order to obtain an estimate of the stability constant of [Fe(2’)], we determined the ,competitive equilibrium constant with EDTA:I4 [Fe(2’)]

+ EDTA4- == [Fe(EDTA)]- + 2‘

Using the 3:l DMF-water pH 8.5 buffer, we measured the decrease in the optical density of the spectrum of [Fe(2’)] (Figure 4) for 1: 1 and 2: 1 Na,EDTA: [Fe(Z’)] ratios. Equilibration was complete after about 5 h at 25 “C. For both of these ratios the effective stability constant, pKeff,of [Fe(2’)] was found to be 24.3, assuming that the stability constant of [Fe(EDTA)]- and the

Synthetic Iron(II1)-Specific Chelators proton dissociation constants of EDTA are the same for our solvent mixture as they are in water.I5 Similarly, we assumed that the proton dissociation constants of 2 would be similar to those observed for other polyhydroxamic acids in water and that we could use these values for our solvent system. Because of the aqueous insolubility of 2, we were unable to determine its proton dissociation constants, but for other hydroxamic acids these constants are in the rangeI4 of 8-10. The proton interference term,I6 aL, for 2 is dominated by the values of its highest dissociation constant at pH 8.5; the other two constants contribute insignificantly to the complexation constant. The highest acid dissociation constants of polyhydroxamic acids vary between 9.5 and 10. Using 9.5 as the highest constant for 2, we find a stability constant for [Fe(2')] of 27.3, and using a value of 10 gives a stability constant of 28.8 for the complex. The stability constants found for [Fe(2')] are within the range expected for a sexidentate hydroxamate complex of iron( 111) despite the assumptions we have made. Thus the biological hydroxamic acid siderophores for iron(II1) complexes have stability constants in the range14 of 29-32.5, that of iron(II1) tris(acetylhyroxamate) being" 28.3.

Discussion Although the hydroxamic acid analogue 2 is specific for iron(111) and forms very stable iron(II1) complexes, its value as a drug appears to be limited by the water insolubility of both ligand and its iron(II1) complex. Presumably these problems could be circumvented by, for example, attaching sulfonic acid groups to the benzyl rings. The main purpose of this paper, however, is to provide a reasonably simple methodology for the synthesis of sexidentate hydroxamic acid ligands analogous to 2. The approach outlined here appears to be less complicated than a recently described strategy reporteds for a similar hydroxamic acid ligand.

Experimental Section A. Platform 11. (1) 1,3,5-Tris(bromomethyl)benzene (10). To 1,3,5-tris(hydroxymethyl)benzene (9) (10.0 g, 0.06 mol; prepared by a LAH reduction of triethyl-1,3,5-benzenetricarboxylate(8) using the

method of Safar',) cooled in an ice bath under N, was slowly added PBr, (52.8 g, 0.195 mol). The resultant stirred mixture was slowly allowed to come to room temperature and was then heated at 11 5 OC for 1 h. After cooling to room temperature, the orange solution was quenched with ice water and the mixture was extracted with CH2CI2. The combined organic extracts were washed with cold saturated aqueous NaHC 0 3 and water and then were dried over MgSO,. The solvent was removed under reduced pressure, and the residue was crystallized from hot hexane, yielding 10 (17.4 g, 82%; mp 92-93 "C) as a white solid. 'H N M R (CDC13): 6 4.5 (s, 6 H), 7.4 (s, 3 H). Anal. Calcd for C9H9Br3: C, 30.3; H, 2.5; Br, 67.2. Found: C, 30.5; H , 2.6; Br, 66.7. (2) 1,3,5-Tris(aminomethyl)benzene (11). Freshly crystallized 10 (25.0 g, 0.07 mol) in dry D M F (60 mL) was added over 20 min to a stirred suspension of potassium phthalimide (40.9 g, 0.221 mol) in dry D M F (200 mL) maintained at 120 OC under N2. After 20 min at 120 OC, the reaction was cooled slightly, water (150 mL) was added, and then this mixture was poured into vigorously stirred water (3 L). The mixture was filtered, and the precipitate, after being washed with water and Et20, was air-dried, yielding 1,3,5-tris(phthalimidomethyl)benzene (36.4 g, 94%). Anal. Calcd for C33H2,N306:C, 71.4; H, 3.8; N, 7.6; 0, 17.3. Found: C, 71.2; H, 3.8; N , 7.6; 0, 17.6. A solution of hydrazine hydrate (85%, 12.2 g 0.22 mol) in EtOH (95%, 20 mL) was added to a suspension of 1,3,5-tris(phthalimidomethyl)benzene (36.4 g, 0.065 mol) in EtOH (95%, 750 mL), and the vixture was refluxed for 12 h. The reaction was cooled slightly and was then carefully acidified (pH -3) with HCI (12 M). When it had come to room temperature, the mixture was filtered and the precipitate was washed well with 95% EtOH. The combined filtrates were evaporated to dryness under reduced pressure, and the white residue was slurried with water and filtered out. The aqueous filtrate was evaporated to dryness under reduced pressure, and the residue was crystallized from H20-EtOH-acetone to yield long white needles of 11 as the trihydrochloride (16.4 g, 91%; mp >300 "C). 'H NMR (D,O): 6 4.4 (s, 6 H), 7.6 (s, 3 H). Anal. Calcd for C9HI8CI3N3:C,

(19) Safar, M.; Kaka, Z.; Galik, U.; Landa, S. Collect. Czech. Chem. Commun. 1974, 39, 3390.

Inorganic Chemistry, Vol. 26, No. 3, 1987 399 39.4;H,6.6;C1,38.7;N, 15.3. Found: C,39.1;H,6.7;CI,38.6;N, 15.1. B. Catecholate Arm. 2,2-Dimethyl-1,3-benzodioxole-4-carbonyl Chloride (6). Catechol was converted to 2,2-dimethyl-l,3-benzodioxole (4) by the method of S l o ~ f fand , ~ ~4 was then carbonylated according to the method described by Carey,,' yielding 5. A solution of 5 (8.92 g, 0.046 mol) in dry CH2CI2 (100 mL) under N, was cooled to 0 OC. Triethylamine (7.68 mL, 0.055 mol) and 1 drop of dry DMF were added, followed at 0 OC by the slow addition SOCI, (5.0 mL, 0.069 mol). The yellow solution was kept at 0 OC for 3 h and then at room temperature for 1 h. After the volatiles were removed under reduced pressure, the residue was triturated with hot hexane. The hexane was removed under reduced pressure and the residue crystallized from hexane to yield 6,' ((7.62 g, 78%). C. Hydroxamic Acid Arm. (1) O-Benzoyl-N-benzoyl-N-(2carboxyethy1)hydroxylamine (15). A solution of 3-(hydroxyamino)propionic acid (14)(5.0 g, 0.0475 mol; prepared from the Michael adduct 13 of (E)-benzaldoxime22and methyl acrylate by the method of Bellasio2)) and NaOH (1.9 g, 0.0475 mol) in water (15 mL) under N2was ice-cooled and was vigorously stirred while benzyl chloride (1 1.05 mL, 0.095 mol) in dry E t 2 0 (30 mL) was added over 10 min. After 15 min NaOH (1.9 g, 0.0475 mol) in water (15 mL) was added dropwise and after an additional 10 min more NaOH (1.9 g, 0.0475 mol) in water (15 mL) was added dropwise to the vigorously stirred reaction. The mixture was then stirred 10 min with ice-cooling and 0.5 h at room temperature. Ethyl acetate (200 mL) was added, and the layers were separated. The organic layer was washed with cold saturated aqueous N a H C 0 3 (40 mL); the combined aqueous layers were acidified (pH 3) with HCI (1 M), and the mixxture was then extracted with ethyl acetate. The organic extracts were washed with brine and were dried over MgS04. The solvent was removed under reduced pressure, and the residue was crystallized from Et20-hexane, yielding 15 (8.5 g, 62%; mp 84.0-85.5 "C) as a white crystalline solid. 'H N M R (CDCI,): 6 2.5-2.9 (m, 2 H), 4.0-4.4 (m, 2 H), 7.0-7.9 (m,10 H), 10.9 (s, 1 H). Anal. Calcd for CI7HISNOS:C, 65.2; H, 4.9; N , 4.9. Found: C, 65.3; H, 4.7; H , 4.4. ( 2 ) O-Benzoyl-N-benzoyl-~-(2-((~-nitrophenoxy)carbonyl)ethyl)hydroxylamine (16). N,N'-Dicyclohexylcarbodiimide(6.21 g, 0.0301 mol) was added to a solution of p-nitrophenol (4.77 g, 0.0343 mol) and 15 (9.0 g, 0.0287 mol) in ethyl acetate (10 mL) under N2. After for 2.5 h of stirring, the mixture was filtered and the precipitate was washed with a little ethyl acetate. The combined filtrates were evaporated to dryness under reduced pressure, and the residue was crystallized from hot EtOH (95%) to give 16 (10.7 g, 86%; mp 96.5-97.5 "C) as white needles. 'H N M R (CDCI3): 6 2.9-3.3 (m, 2 H), 4.2-4.6 (m, 2 H), 7.0-8.2 (m, 14 H). Anal. Calcd for C23H18N207:C, 63.6; H, 4.2; N, 6.5. Found: C, 63.7; H, 4.2; N , 6.3. D. Ligand 1. Freshly recrystallized ll.3HC1 (2.98 g, 0.01 1 mol) in water (100 mL) under N, was neutralized with NaOH (1.32 g, 0.033 mol); CH2Cl (100 mL) was then added, and the vigorously stirred mixture was ice-cooled. Solutions of Et3N (5.03 mL, 0.036 mol) in CH,Cl, (50 mL) and freshly prepared 6 (7.62 g, 0.036 mol) in CH2C12(50 mL) were simultaneously added dropwise to the ice-cooled, vigorously stirred mixture. The reaction was slowly allowed to come to room temperature. After 18 h the layers were separated and CH2CI2was removed under reduced pressure. The oily residue was dissolved in benzene, the mixture was washed with water, saturated aqueous NaHCO,, water, a pH 4 buffer, and water and was then dried over Na2S04. Removal of the solvent under reduced pressure yielded an oily residue from which was crystallized (benzene-hexane) 17 (7.0 g, 83%; mp 82-83 "C) as white needles. These needles contain a molecule of benzene solvate, which was removed at 30 OC under high vacuum. 'H N M R (CDCI,): 6 1.7 (s, 18 H), 4.6 (br d, J = 6 Hz, 6 H), 6.8-7.4 (m, 15 H). Anal. Calcd for C3,H3,N,O9: C, 67.5; H, 5.8; N , 6.1; 0, 20.8. Found: C, 67.5; H, 5.6; N, 6.4; 0, 21.1, A solution of trifluoroacetic acid and water (4: 1, 15 mL) was added to (desolvated) 17 (1.4 g, 0.002 mol) at room temperature under N2,and the mixture was swirled to effect dissolution. On standing, the light yellow solution deposited an off-white solid. After 24 h the mixture was filtered and the precipitate was washed with water and air-dried. The crude product was recrystallized from ethyl acetateEt20-hexane to yield 1 (0.9 g, 79%; mp 123-125 O C (lit9 mp 130 " C ) ) as white needles. 'H N M R (Me2SO-d,): 6 4.5 (br d, J = 6 Hz, 6 H), 6.4-7.4 (m,12 H), 9.1-9.5 (m, 6 H), 12.0-13.00 (br s, 3 H). Anal. Calcd for C30H27N309:C, 62.9; H, 4.8; N , 7.3. Found: C, 62.5; H, 5.2;

-

(20) Slooff, G. R e d . Trau. Chim. Pays-Bas 1935, 54, 995. (21) Corey, E. J.; Hurt, S. D.Tetrahedron Lett. 1977, 3923. (22) Furniss, B. S., Hannaford, A. J., Rogers, U., Smith, P. W. G., Tatchell, A. R., Eds. Vogel's Textbook of Practical Organic Chemistry, 4th ed.; Longman: New York, 1978; p 810. (23) Bellasio, E.; Parravicini, F.; LaNoce, T.; Testa, E. Ann. Chim. (Rome) 1968, 58, 413.

Inorg. Chem. 1987, 26, 400-405

400

N , 7.4. E. Ligand 2. Triethylamine (2.0 mL, 0.0143 mol) was added to a stirred solution of lls3HC1 (1.3 g, 0.0047 mol) in water (10 mL) under N,, and after 5 min 16 (6.3 g, 0.0145 mol) in T H F (80 mL) was added all at once. There was an immediate formation of a bright yellow twophase mixture. After 1 h the cloudy mixture was partially cleared by the addition of water (2 mL) and T H F (20 mL), and after an additional 5 h more Et3N (2.0 mL, 0.0143 mol) was added. The mixture was stirred for a total of 20 h, and then the T H F was removed under reduced pressure. The residue was extracted with ethyl acetate (2 X 75 mL), and the combined organic layers were washed with water (5 X 120 mL) and brine (4 X 100 mL) and were dried over Na2S04. After the extract was passed through a short column of Florisil, ethyl acetate was removed under reduced pressure, yielding 16 (6.0 g) as a light green foam. This was then dissolved in methanolic ammonia (65 g, 8%, w/w) and the mixture was allowed to stand at room temperature for 1.5 h. The bright yellow, slightly cloudy solution was filtered through Celite, and the solvent was removed under reduced pressure, yielding a bright yellow oil, which was taken up in hot EtOH (25 mL). Upon standing (10 h at room temperature, 2 days at 4 "C, and 1 day at -20 "C), a yellow solid was deposited, which was recrystallized twice from hot EtOH, yielding 2 (1.76 g, 50%; mp 162-165 "C) as an off-white solid. 'H N M R Me2SO-d6): 6 2.4-2.8 (m, 6 H), 3.7-4.1 (m, 6 H), 4.2 (d, J = 5 Hz, 6 H), 7.0 (s, 3 H), 7.2-7.7 (m.15 H), 8.4 (br t, J = 5 Hz, 3 H), 9.8 (br s, 3 H). Anal. Calcd for C39H,2N60,: C, 63.4; H , 5.7; N , 11.4. Found: C, 63.2; H , 5.8; N, 11.3. F. Iron(II1) Complexes. A solution of Fe(NH4)(S04)2.12H20 (0.01557 g, 3.23 X mol) in water (2 mL) was treated with 1 (0.01857 g, 3.23 X mol) in MeOH (2 mL). The purple milky solution was made up to 100 mL with a pH 8.50 borax buffer, producing a clear violet solution that displayed absorptions (Varian 2300 spectro-

photometer) at A,, = 498 nm (e = 4697 M-' cm-I) and A,, = 328 nm ( t = 17900 M-' em-'). The spectrum showed no change over several weeks. A solution of Fe(NH4)(S0&.12H20 (0.01205 g, 2.50 X mol) in water (2 mL) was treated with 2 (0.01847 g, 2.50 X mol) in DMF (10 mL), producing a clear red solution. D M F (65 mL) and a pH 8.50 borax buffer (22 mL) were added to this solution in portions, with sufficient DMF being added before the buffer to prevent the appearance of a fluffy orange-red precipitate. The total volume was made up to 100 mL with the buffer, producing an orange-red solution that exhibited only a broad absorption at A,, = 428 nm (e = 3466 M-l cm-I). The absorption reached its maximum intensity after -20 h and remained unchanged for at least 5 days. Stability Constants. For the determination of the stability constant of the iron complex, the same procedure was followed except that Na2EDTA.2H20 (1 or 2 equiv based on 2) dissolved in a little of the buffer was added after iron and the ligand 2 had been mixed. DMF and the rest of the buffer were then added as before. The absorption spectra showed that the intensity of the absorption at 428 nm in the case of 1 equiv of added Na2EDTA.2H20 reached a maximum value after 5 h, a showing 23.8% loss of intensity. For the 2 equiv of added Na,EDTA. 2 H 2 0 the maximum value was reached in under 5 h, showing a 35.0% loss of intensity. Both of these solutions deposited a precipitate on standing after -36 h. Instrumentation. The 'H N M R spectra were obtained on a Varian T-60 spectrometer, and the UV spectra were run on a Varian 2300 spectrophotometer.

Acknowledgment. This work was supported by grants from t h e N a t u r a l Sciences and Engineering Council of C a n a d a , t h e A t kinson Foundation, a n d t h e J. P. Bickell Foundation.

Contribution from the Department of Chemistry, Princeton University, Princeton, New Jersey 08544

Ultraviolet Resonance Raman Study of Deoxyguanosine Monophosphate and Its Complexes with cis-(NH3)Pt2+,Ni2+, and H+ Joseph R. Perno, Dory Cwikel, and Thomas G. Spiro* Received May 23, 1986 Raman spectra are reported with deep ultraviolet excitation at 240 and 218 nm for dilute (5 mM) aqueous solutions of 2'deoxyriboguanosine 5'-monophosphate (dGMP), its N7-protonated form (pH I S ) , its 1: I complex with aqueous Ni2+.and its 1: 1 and 2:l complexes with ~ i s - ( N H ~ ) ~ p The t ~ * .C=06 stretching band at 1677 cm-' in dGMP shifts up to 1702 cm-l on N7 protonation and to 1686 cm-l in the 2:1 Pt complex, reflecting the polarizing effect of these positive centers. In the Ni complex, however, this band shifts down to 1660 cm-I, an effect attributed to intramolecular H bonding of the carbonyl oxygen atom by Ni-bound H20, as observed in the crystal structure. In the 1:l Pt complex, a smaller downshift is seen, to 1670 cm?, and is suggested to be the resultant of opposing effects of polarization and intramolecular H bonding via the Pt-bound H,O molecule; this H 2 0 is displaced by another guanine N7 in the 2:l complex. Several guanine ring modes show progressive increases or decreases in frequency and/or intensity upon binding Ni, Pt, and H* due to polarization. Guanine modes at 682 and 855 cm-' in dGMP, which are known to be sensitive to the sugar conformation, develop prominent shoulders at 666 and 864 cm'' in the Ni and Pt complexes, and also in riboguanosine 5'-monophosphate (rGMP). These additional components are attributed to molecules with a C3'-endo ribose configuration, while the 682- and 855-cm-' bands are associated with the C2'-endo conformation, which is dominant in dGMP. The crystal structure of the Ni complex shows a C3'-endo conformation, which allows the 5'-monophosphate linkage to accept H bonds from Ni2+-boundH 2 0 molecules. It is suggested that this structural feature is only partially retained in solution. Prominent 666-cm-I C3'-endo bands are also seen for Pt (but not Ni) complexes of 3'-dGMP and 2'-deoxyguanosine. Pt binding is suggested to stabilize the C3'-endo conformation via an electronic polarization effect.

Introduction has recently become possible t o obtain R a m a n spectra with deep ultraviolet excitation by using pulsed YAG lasers and various frequency shifting devices.'" This technology has made it possible to explore resonance effects in Raman spectra of simple aromatic chromophores, including aromatic side chains of proteins5-'* a n d the purine and pyrimidine bases of nucleic acids.l3-I6 Rich spectra of purine a n d pyrimidine ring modes have been reported, with t h e e n h a n c e m e n t p a t t e r n varying strongly with laser excitation wavelength, a s different R-R* excited states a r e accessed. V a r iable-wavelength excitation c a n therefore b e used t o sort out and

*To whom correspondence should be addressed.

0020-1669/87/1326-0400$01.50/0

identify t h e different vibrational modes, whose overlap tends to obscure t h e b a n d assignments in nonresonance R a m a n spectra (1) Ziegler, L. D.; Hudson, B. S. J . Chem. Phys. 1981, 74,982. (2) Ziegler, L. D.; Hudson, B. S. J . Chem. Phys. 1983, 79, 1134. (3) Asher, S. A.; Johnson, C. R.; Murtaugh, J. Rea. Sci. Instrum. 1983,54,

1657.

(4) Fodor, S. P. A,; Rava, R. P.; Copeland, R. A,; Spiro, T. G . J . Raman

Speclrosc., in press.

( 5 ) Rava, R. P.; Spiro, T. G. J . Am. Chem. Sac. 1984, 106, 4062-4064. (6) Johnson, C. R.; Ludwig, M.; ODonnell, S.; Asher, S. A. J . Am. Chem. SOC.1984, 106, 5008-5010. (7) Rava, R. P.; Spiro, T. G. J . Phys. Chem. 1985, 89, 1856-1861. (8) Rava, R. P.; Spiro, T. G. Biochemisrry 1985, 24, 1861-1865. (9) Copeland, R. A.; Dasgupta, S.; Spiro, T. G. J . Am. Chem. Sac. 1985, 107, 3370-3371.

0 1987 American Chemical Society